Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Most brilliant novelty from cct7 Witchess-Arasan

Author: Stephen A. Boak

Date: 10:57:32 02/15/05

Go up one level in this thread


On February 15, 2005 at 11:21:56, Vincent Diepeveen wrote:

>
>I hope that you realize 750 rating points means roughly 0% chance.
>
>What is the chance in YOUR opinion a program without book in 2005 has to win a
>world champs event?
>
>If you say 0%, that means 700+ rating points.
>
>Vincent

Your statistics are somewhat flawed.

1. A single program vs. another single program involves an ELO difference which
may be translated into a scoring probability or expectation.  Or a scoring
probability or expectation (based on experimentation) may be translated into
that ELO difference.  More or less.

2. The probability of winning a tournament with multiple opponents involves
multiple rounds and multiple cross-probabilities between a single program and
the *many* programs that are opponents.

For any arbitrarily chosen ELOs, the probability of winning a tournament is
generally far less than the probability of scoring for a comp vs. comp single
game match.

2A. Example, if 100 ELO points means approx. a 2-1 scoring advantage for the
higher program (stronger program has a winning expectancy of approx. 60-65%),
the lower program's expected scoring vs. that single program is approx. 35-40%.

2B. However, if that program must play a field of 9 other opponents each of
which outrates the program by 100 points, the chances of the low ELO program
winning a multi-round tournament against 9 players is extraordinarily low (vs.
the 1-on-1 scoring ration of 30-35%).  The chances of winning such a tournament
may easily approach zero long before the ELO gap gets anywhere near to 700
points.

Regards,
--Steve






This page took 0.04 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.